Fog computing task scheduling optimization based on multi-objective simplified swarm optimization

Abstract

With the dramatic growth of data volume, the cloud computing structure has faced a severe difficulty, the latency. To deal with this problem, researchers have proposed the fog computing structure, which can successfully release the computation loads from one datacenter of cloud to multiple local fog devices. Hence, the tasks will be processed at the local fog device and avoid transmitting to datacenter which is not cost-effective, and the results can be transmitted to the users immediately. The main differences of task scheduling between cloud computing and fog computing are the processors specifications, such as processing rate, cost per unit time. The processing rate of processors in datacenter is fast but the cost per unit time is costly. However, the processing rate of fog devices is not as fast as datacenter but it is relatively cost-effective. Consequently, we are facing a multi-objective optimization problem. Hence, we adopted an elite MOSSO (Multi-Objective Simplified Swarm Optimization) with considering the characteristics of fog computing paradigm.